FABRACADABRA!

Fan created remixes and rare variations
Josh
Posts: 60
Joined: Sun Feb 21, 2021 5:57 am

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by Josh »

BTL1210 - You're welcome, those were mostly just proof of concept to see if AI remastering could enhance these demixes even further. In your opinion, would it be worth it to apply the AI mastering across the board?
DVWerks
Posts: 19
Joined: Thu Feb 25, 2021 8:08 pm
Been thanked: 7 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by DVWerks »

Josh wrote: Sat Jan 14, 2023 2:40 pm Hey LR, I have a silly question for you. Have you ever tried using AI mastering software in addition to AI demixing? I ran a few of your demixes through aimastering.com and I'm super impressed with the result.

Here are Day Tripper, And Your Bird Can Sing, and Strawberry Fields Forever with AI mastering applied. hxxps://we.tl/t-EfSKjl0evY
Is the AI mastering through a particular software or website, and is it adjustable? I checked out these samples and Strawberry Fields seemed particularly aggressive, with a lot of the track waveform looking brickwalled. Is the AI mastering trying do anything other than maximize loudness?
User avatar
applebonkerz
Posts: 159
Joined: Tue Aug 10, 2021 11:52 pm
Has thanked: 222 times
Been thanked: 14 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by applebonkerz »

Why in the world would Lord Reith, or anyone else for that matter, want to ruin his fine work that so much time has been put into? One of the reasons these sets still sound so excellent is precisely because they haven't gone through ridiculous "remastering" to suck all the subtleties out of them! :roll:
User avatar
Lord Reith
Posts: 4608
Joined: Thu Feb 18, 2021 8:22 am
Location: BBC House
Has thanked: 140 times
Been thanked: 3972 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by Lord Reith »

Josh wrote: Sat Jan 14, 2023 2:40 pm Hey LR, I have a silly question for you. Have you ever tried using AI mastering software in addition to AI demixing? I ran a few of your demixes through aimastering.com and I'm super impressed with the result.
I'm familiar with AI mastering and there was a big thread about it on BZ a few years ago with lots of people trying their hand at it. But personally I don't care for it since it tries to make the mix sound like a 2023 master, and I want it to sound like a 1960s master.

No objection to you posting examples but I would rather you not do everything. That would just get confusing for people.
Women there don't treat you mean, in Abilene
User avatar
Tex
Posts: 1157
Joined: Fri Feb 19, 2021 1:12 am
Location: Texas
Has thanked: 6 times
Been thanked: 586 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by Tex »

AI remastering? Meh. I think most people here have a DAW program so they can remaster/compress to their heart's content. You either have to have a sound system adjusted to perfect playback OR a sound system simply adjusted to good reference material. My stereo was always adjusted so Eagles, Steely Dan and Pink Floyd sounded good. You don't want to compete with modern pop/rap.

Audacity is free and I use that for opening files and spectrum checking.

I use an old copy of "Magix Audio Studio" (check Ebay) which is a real consumer DAW and has very nice sounding 3-band compressor. If I want a NUKE-U-LAR compress and Modern EQ there's an option in Magix Audio Cleaning Lab but nothing really deserves such abuse.

I "remastered" a few things posted here but it would be silly for me to post. I would only ever post my own remixes anyways even if it means essentially copying other's work.
User avatar
Lord Reith
Posts: 4608
Joined: Thu Feb 18, 2021 8:22 am
Location: BBC House
Has thanked: 140 times
Been thanked: 3972 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by Lord Reith »

This all raises an interesting point which I'm not sure I've ever commented on before but I will now since it seems kind of relevant.

I'm constantly seeing audiophiles or remixers talking about "A-B-ing" two different sources to decide which is the "best". Leaving aside for a moment that "best" is a completely subjective concept, that is the exact wrong way to do it. Listening to two differing sources one after the other is not an aid to determining quality, it is a hindrance. You need to "clean the palette" first, either by taking a break for a few minutes or lsitening to something else entirely unrelated. Then come back and listen to the B source.

The best way I can explain that is to invoke The Pepsi Challenge. Those of us ancient enough to remember will recall this famous TV promotion which pitted Pepsi and Coke against each other in a blind A-B taste test. Pepsi, surprisingly, almost always won. And yet that didn't translate into much bigger sales somehow. When people went back home, they seemed to go back to prefering Coke. Eventually someone worked out that people will nearly always go for the sweeter taste in an A-B test, and Pepsi just happened to have a little more sugar. But that didn't mean Pepsi was better, or that people would prefer to drink it over Coke on a daily basis.

It's the same with audio. In an A-B test people will nearly always prefer the source with more treble, bass and volume. It seems to have more impact, and it makes the other source sound limp. But that doesn't necessarily mean that in day to day listening you're going to like the bigger and bolder version better.

I'm not saying that fast A-B-ing of two different sources is not useful sometimes, so long as there is something quantifiable you are trying to measure, like which source is brighter or which is more noisy. But as a determiner of "what sounds best" you have to be aware of your brain's tendency to be tricked. The ears relish treble like the tongue relishes sweetness. But just as something overly sweet can become sickly, something with too much treble, bass or compression can quickly become fatiguing. You'll notice this when listening to something and after 15 or 20 minutes your brain seems to faze over and the music loses its enjoyability.

Unfortunately, the streaming platforms and radio stations are engaged in their own Pepsi Challenge, and it means that if your music is not mastered in the same big-sounding way then it sounds very poor next to the competition. But it is not creating better sounding albums.
Women there don't treat you mean, in Abilene
User avatar
applebonkerz
Posts: 159
Joined: Tue Aug 10, 2021 11:52 pm
Has thanked: 222 times
Been thanked: 14 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by applebonkerz »

Truth. Amen.
BTL1210
Posts: 4778
Joined: Mon Aug 30, 2021 10:27 am
Has thanked: 3 times
Been thanked: 233 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by BTL1210 »

Amen.
Appleyard
Posts: 102
Joined: Sat Sep 25, 2021 9:46 am
Has thanked: 13 times
Been thanked: 26 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by Appleyard »

Wow. Someone has finally managed to put in words what I (and many, many others) have been feeling for over 2 decades. Thank you Lord Reith. That Pepsi challenge explains it perfectly. I guess this is why many of us go back to 80s pressed CDs for some albums rather than listen to modern remasterings.
Also, I have loved your tweaks on these Fabracadabras. Just finished listening to them all in the car which is where I find I can really appreciate your instrument placement without headphones. The last fortnight's journeys to work have been wonderful!
I have a request ( :D !) and I wondered if it were possible for you to have a go at it. In Strawberry Fields Forever, I've always loved both of the different pans in the 2 stereo mixes. Would you be able to create a mix that had the US strings pan on the "coz I'm going to" and the UK swarmandal pan? I didn't hear the Capitol mix until later in life when I started collecting and I love how it covers up that edit point perfectly. However, I grew up with the swarmandal pan and so of course I love that too! Would be marvellous to have a definitive version. My attempts have failed!
Anyway, I doff my hat to you, m'lud for your fine work.
Many, many thanks.
Mixerrog
Posts: 279
Joined: Sat May 22, 2021 5:01 pm
Has thanked: 41 times
Been thanked: 100 times

Re: FABRACADABRA! (V1,2,4 new! V3, V6, V7, V8, V9)

Post by Mixerrog »

Here is my comments about source material that I use --

First, I check the waveform to find a song file with the least loudness applied & best dynamic range which can easily be seen in the waveform view.

Second, I check the spectrum to make sure it if a full frequency without dropouts. If a record transfer, I usually apply a steep filter at about 20 HZ to eliminate turntable rumble which can be very hard on low frequency speakers at loud volumes due to excessive needless cone movement.

Third, this is my listening test on headphones usually for good quality sound. I am not overly concerned with EQ as I can apply or match that usually in a Spectrum view after remixing the song.

I find that loudness compression or brick walling as I call it, can really hurt the quality of stem separations using AI or make it much harder to do
manual spectral editing due to the difference in brightness which represents audio level in the spectrum view is much harder to see so harder to work with.

Rog
Post Reply